Contents lists available at ScienceDirect

Remote Sensing Applications: Society and

Remote Sensing Applications: Society and Environment 1 (2015) 72–84

http://d2352-93

n CorrE-m

journal homepage: www.elsevier.com/locate/rsaseEnvironment

Current situation and needs in man-made and natech risksmanagement using Earth Observation techniques

Sabina Di Franco n, Rosamaria SalvatoriIIA – CNR, Rome, Italy

a r t i c l e i n f o

Article history:Received 23 March 2015Received in revised form9 June 2015Accepted 10 June 2015Available online 29 July 2015

Keywords:NatechMan-made hazardsIndustrial accidentRisk managementPreparednessEmergencyRecoverySmall satelliteUAV

x.doi.org/10.1016/j.rsase.2015.06.00485/& 2015 Elsevier B.V. All rights reserved.

esponding author.ail address: difranco@iia.cnr.it (S. Di Franco).

a b s t r a c t

The Earth Observation (EO) techniques are becoming increasingly important in risk man-agement activities not only for natural hazards and natural disaster monitoring but also toride out industrial and natech accidents. The latest developments in the aerospace industrysuch as sensors miniaturization and high spatial and temporal resolution missions, devoted tomonitoring areas of specific interest, have made the use of EO techniques more efficiently andare vready to be used in near real time conditions. This paper summarize the current state ofknowledge on how EO data can be useful in managing all the phases of the Industrial/natechdisaster, and from the environmental conditions before the accident strikes to the postaccident relief, from the scenario setting and planning stage to the damage assessment.

& 2015 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732. State of the art of the use of EO for industrial and natech risk management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

2.1. Explosions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.2. Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.3. Nuclear accidents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 782.4. Oil spills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

3. Future development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 803.1. Small satellites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813.2. UAV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 73

1. Introduction

Risk management is a complex activity that requires amultidisciplinary approach. When a disaster occurs, everyminute is crucial to save lives, protect people, property andthe environment and to react in a coordinated and con-scious way which makes the real difference between asuccessful emergency management and failure. The eventscaused by disasters are somehow repetitive and form acycle that can be divided in four phases: mitigation andpreparedness (before the catastrophe strikes); responseand recovery – reconstruction included – that occur afterthe disaster. The mitigation phase consists of all actionsneeded to reduce the impact of future disasters (Menoniet al., 2012). These can be divided in structural (technicaland structural solutions) and non-structural measuressuch as land use-planning, legislation measure and eva-cuation planning (Galderisi et al., 2008). Preparednessphase comprises the actions taken to reduce the impactswhen the disaster is forecast or imminent. Response per-tains to actions taken during and immediately after thedisaster, with the main aim to save and safeguard humanlives. The term recovery refers to the process of restoringservices and repairing damage after the disaster has struck(Alexander, 2002).

Keeping in mind this cycle the contribution of the sci-entific community and the use of innovative technologiessuch as those related to Earth Observation are of strategicimportance during all the phases of the emergency man-agement (Joyce et al., 2009). The emergency managementplanning can be considered similar to an urban or regionalplanning process; both require that the local conditionsand geographic characteristic of the place are properlyconsidered, especially in term of hazardousness (Alex-ander, 2006).

Fig. 1. Number of events and Industry (2002–2012). eMARS JRC

Moreover, the crisis events are often characterized byrapid evolutionary dynamics, with scenarios that can oftenchange significantly in a very short time. Therefore, betteremergency management necessarily passes through thequality and quantity of observations and information, aswell as the speed at which the information can be trans-ferred and made clear and usable by decision makers.

The industrial risk, from a risk classification point ofview, can be considered as a part of the wide category ofman-made hazards. The man-made hazards, with somevariations depending on different classifications, include:technological hazards, nuclear risk, transport risk andother anthropic activities such as business, infrastructureand technological networks management, that can be asource of danger to humans and the environment (AA. VV.,2006); in the man-made hazards perspective the envir-onmental risk is related to the probability of an eventcaused by unexpected alteration of physical and chemicalparameters in the environment (water, air and/or soil),that have immediate or short-term effects on the health ofthe resident population. Another definition, used in tech-nical papers, highlights the difference between “human-made disaster” that are caused directly by human activitiesand “human-induced disaster”, natural disaster that areaccelerated/aggravated by human influence (Van Westen,2002).

In this heterogeneous framework of hazards, risks andevents, some significant industrial accidents are known tobe caused or triggered by natural disasters. In the inter-national literature, this type of accident is defined asnatech or "Natural-Technological" event. One of the natechdefinitions recite as follows: "Technological accidents, likefires, explosions and toxic releases that may occur inindustrial complexes and along the distribution networkas a result of natural disasters of natural matrix" (Clerc andLe Claire, 1994; Lindell and Perry, 1996; Cruz et al., 2004).

– European Commission, Major Accident Hazards Bureau.

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8474

The natech scenarios are emerging scenarios and theyare considered to become more frequent due to the effectsof climate change (Krausmann et al., 2011, Salzano et al.,2013). National authorities have to identify those areasthat may be affected by such events (European Commis-sion - Joint Research Centre, 2004). Currently, identifica-tion and mapping of these hazards is not very commonand the Joint Research Centre (JRC – European Commis-sion) is developing a tool for mapping and rapid assess-ment of this type of emergency, RAPID-N, which operatesat a global scale (Girgin and Krausmann, 2013).

To investigate, monitor and analyse the industrialaccident in Europe, the European Commission JRC (JointResearch Centre) was established in 1982 by the EU'sSeveso Directive 82/501/EEC, the Major Accident ReportingSystem (MARS and later renamed eMARS). The purpose ofthe eMARS is to facilitate the exchange of lessons learnedfrom accidents and near misses involving dangerous sub-stances in order to improve chemical accident preventionand mitigation of potential consequences. “MARS containsreports of chemical accidents and near misses provided tothe Major Accident and Hazards Bureau (MAHB) of theEuropean Commission's Joint Research Centre from EU,OECD and UNECE countries (under the TEIA Convention)(Fig.1). Reporting an event into eMARS is compulsory forEU Member States when a Seveso establishment isinvolved and the event meets the criteria of a “majoraccident” as defined by Annex VI of the Seveso III Directive(2012/18/EU, Major Accident Reporting System https://emars.jrc.ec.europa.eu/).

In many cases the work of rescue may be delayedbecause the rescuers cannot reach the affected areas dur-ing the incident nor immediately after the incident, butonly after they have assessed the conditions to operatein safety conditions (accident in a central nuclear or

Fig. 2. Number of events per type of industrial accident reported from

accidents with leakage of toxic gases). In case of release oftoxic substances in the air, rescue teams must quicklyevacuate residents from the area considered at risk whosesize must be defined, from the location of the source ofemission, the direction and velocity of the wind and fromother meteorological conditions at the time of the acci-dent. It must be taken into account the location of theemergency areas relatively to geographical characteristics,the roads network and the presence of important infra-structures (dams, bridges, stations, public buildings).These considerations are to be performed effectively andquickly getting each time a precise and effective feedbackof data collected on-site.

As reported to the Major Accident Reporting System, inrecent years (2000–2012) industrial accidents in Europewere about 490:201 releases of toxic substances, 153 fires,132 explosions and 5 accidents occurred during goodstransportation (eMARS) (Fig. 2).

2. State of the art of the use of EO for industrial andnatech risk management

Within the emergency management framework, EarthObservation (EO) systems could play a key role, givingtimely and accurate information not only on the extensionand degree of damages, but also on post event emergencyactivities.

Nowadays the satellites orbiting around our planet areequipped with active and passive sensors that operate overthe entire spectral range of wavelengths from ultraviolet tomicrowave; due to this reason in many emergency situa-tions; it is possible to suppose that most of the earthsurface can be monitored from space at different times.

2000 to 2012 as reported in eMARS JRC – European Commission.

Fig. 3. Field of EO data application for revisit time VS GSD. Modified fromSandau (2010).

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 75

Space missions currently operating are versatile andmulti-purpose; moreover the sensors are dedicated tospecific themes (e.g. Observation of polar ice, vegetation,water quality, etc.). The new sensors data are ideally con-nected to data taken by the first Earth Observation mis-sions from the seventies; this allows to have consistentdatasets comparable with each other. These data allow toperform multi-temporal analyses that until some time agowere impossible to perform (e.g. urban sprawl or extent ofthe polar caps).

Missions of ESA (European Space Agency), EUMETSAT(European Organisation for the Exploitation of Meteor-ological Satellites), NASA (National Aeronautics and SpaceAdministration), NOAA (National Oceanic and AtmosphericAdministration), DLR (Deutschen Zentrum für Luft- undRaumfahrt), together with those of ASI (Agenzia SpazialeItaliana), (Cosmo-SkyMed) provide a very wide variety ofobservational systems, which will be further enriched bythe Sentinels missions, under the new European programCopernicus, just officially started with the launch of Sen-tinel-1 in April 2014. (European Commission, CopernicusEmergency Management Service-www.emergency.copernicus.eu)

Most of the European space programs have focused onemergency management, not only from a technical-sci-entific point of view but also with the investment of sig-nificant resources by industrial companies. For example inthe GMES (Global Monitoring for Environment andSecurity) Program, now renamed Copernicus, emergencymanagement immediately gained a crucial role. This issue,in fact, is among the first Fast Track services provided andfunded by the program and currently one of the six coreservices which is set to become operational in the nextseven years. In 2008 the group Earth Observation of theEuropean Commission has launched the Global EarthObservation System of Systems (GEOSS); in 2008 EuropeanUnion's Space Council always reiterated the need for rapidimplementation of the Global Monitoring for Environmentand Security (GMES) (Aschbacher and Milango-Perez,2012). The program now provides data on demand: from2012, the service has provided maps OT as a result of about100 requests for activation, mainly relating to naturaldisasters such as hydro-meteorological and fire emergency(European Commission. Copernicus Emergency Manage-ment Service www.emergency.copernicus.eu).

Satellite images are useful not only as sources of real-time, or near real-time data to handle hazards, but also astechnologies that can help “to prevent or mitigate theeffects of those hazards” (Showalter and Ramspott, 1999).Industrial accidents, both caused directly by humanactivities or triggered by natural hazards (natech) takeplace, with the exception of oil spills, in previously knownindustrial and urban areas. In comparison with naturalhazards, industrial risks and man-made hazards are less“spatially unpredictable” because they can occur only inareas that are dedicated to industrial activities (Marzoet al., 2015). At the same time natural hazards are lesspredictable but usually take place at much larger scalethan industrial risk and man made hazards that are relatedto single plants or industrial zones. In the last yearssatellite images resolution has greatly increased thus

permitting a close and accurate examination of areas thatneed a higher scale to be monitored.

Table 1 shows the optical (family, resolution and foot-print) and radar sensors that can be used in the case ofnatural events or industrial accidents (Kucera et al., 2012).

In this response to disaster management needs, theoptimal spatial resolution, as well the extension of theregion of the earth surface needed, vary in accordancewith the type of disaster. Sometimes more than a scene isnecessary to analyse affected zones in detail and moreimages with the right spectral resolution must be chosento cover the area.

Within the GEOSS project, currently available satellitesensors have been successfully used for monitoring thedamage caused by natural disasters (earthquakes, floods,forest fires, etc.) and, in some cases, they are also widelyused for prevention activities. Industrial accidents aremore complex to analyze remotely as the area of origin ofthese accidents may be very small but when a long timespan is considered, the damaged area can actually be muchextended.

The following diagram (Fig. 3 modified from Sandau,2010) on the x-axis shows different values of GroundSampling Distance (GSD-resolution) and on the y-axis, thesatellite revisit time (h) and underlines the characteristicsof the images useful in different field of EO data applica-tion, such as mapping, geology, meteorology and disastermonitoring.

To analysis of the territory affected by the calamitousevent with EO data is therefore necessary to make aninformed choice in the type of data to be used, especiallywith regard to their spatial and spectral resolution. Forexample multispectral high resolution images can be usedto obtain land-use maps necessary to derive multi-hazardmap, (Sengupta, 2007) and also to analyze the impact ofman made disaster on impervious regions (Rout et al.,2005). Different information can be derived by infraredimages (both in near and thermal infrared).

Major accidents in industrial plants which result in therelease of hazardous chemicals occur suddenly and couldcause a domino effect, with serious consequences and awide range of damage (Antonioni et al., 2009); it followsthat remote data cannot be used as an activity of "pre-diction". Remotely sensed data, if collected promptly,however, can be an instrument of great importance during

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8476

the first intervention phases, helping to formulate anintervention strategy and to implement the "first aid"activities. These data are undoubtedly relevant for theassessment of damage to built areas so much as to theenvironment (air, water, and soil).

A rapid, accurate and effective assessment of thesituation after the incident is definitely the most impor-tant part of both rescue and recovery operations (Galderisiet al., 2008).

Summing up the effects of an industrial accident can bedivided into three main types: fire, explosion and releaseof toxic and harmful substances (e.g. oil spills, or nuclearaccidents) (El Hajji et al., 2013). The following sub-paragraphs describe the main scenario, the EO techniquesuseful to deal with different type of accident, along withsome real cases of natech events.

2.1. Explosions

Major accidents involving hazardous chemicals, inaddition to the release of pollutants in different environ-mental matrices in the form of emissions, can lead to cri-tical events such as explosions or collapses that easilydegenerate into a series of chain reactions (Antonioni et al.,2007, 2009). The accidents produce fumes and damage tostructures related to the high temperatures involved,causing casualties and huge economic losses. All of theseaccidents are characterized by an unexpected suddennessoften with release of a large amount of toxic substances ofdifferent types (gases, "smoke", powders, liquids) that canspread in the neighboring areas, both natural and urban,altering significantly different environmental matrices(Marzo et al., 2012). The damaged area can cover tens oreven hundreds of square kilometers and rescue operationshave a different complexity in relation to restorationactivities and remediation of environmental damage.

Two relevant examples of natech triggered explosionaccidents are: the earthquake-triggered explosions in thestorage tank farm of Chiba (Japan) refinery, during theGreat East Japan earthquake in 2011 (Krausmann and Cruz,2013), and the explosion caused by lightning in a cerealprocessing factory, at Huesca-Spain-in 2005 (French Min-istry of Ecology 2014 ARIA -analysis, research and infor-mation on accidents – database).

Keeping this situation, it becomes extremely importantto have a "remote" technology to obtain timely informa-tion on the incident, both at macroscopic scale (regionalclassification of the area where the accident occurred) andat extremely detailed scale (level of industrial plant) inorder to monitor the evolution of the phenomenon,especially during particular environmental and meteor-ological conditions. Given the most common accidentscenarios, the current state of the art related to the EO ismainly focused to support the mapping services, includingnear-real time maps, detailed maps, pre and post emer-gency maps and also maps that can allow to quantifyphysical exposure, one of the variables required in disasterrisk assessments (Ehrlich and Tenerelli 2013, Dyke et al.,2010). Particularly in the pre-emergency these tools ofremote sensing can be used to plan, locate industries atrisk, and as expected in the emergency plans, depending

on the risk scenarios assumed (explosion, release, fire) todefine the damage areas, and to verify the exposure(Ehrlich and Tenerelli, 2013) of buildings, critical infra-structure (bridges, dams, roads, power plants, etc.) andpeople potentially affected. After the accident, the map-ping allows an initial damage assessment, supports theemergency management and helps the immediate ver-ification of the entity of the event.

2.2. Fire

The existing satellite sensors have been equipped withchannels in the mid-infrared (3–4 μm), AVHRR (AdvancedVery High Resolution Radiometer)/NOAA, MODIS (Moder-ate Resolution Imaging Spectroradiometer)/AQUA TERRAthat are used to provide data on active fires. The datacollected by these channels have limited spatial resolution(about 1 to 4 km) and may have problems of saturationalready at low temperatures (with the exception ofMODIS). The radiometric values registered in this channel,in addition, also include the contribution of the solarradiation reflected by the earth's surface; it follows that, insome cases, the high signal recorded in this channel doesnot correspond to an active fire and that generates falsealarms and prevents a quantitative characterization oflarge fires. The multi-spectral sensors as Thematic Mapper(TM) or Enhanced Thematic Mapper (ETM) or OperationalLand Imager (OLI), and the Thermal Infrared Sensor (TIRS),on board of Landsat and sensor on board of ASTER/EARTH,with higher spatial resolution, do not have a channel at 3–4 μm, essential for identifying diurnal fire: their channelsare less sensitive to smoldering fires and are more influ-enced by the effects of solar reflection. The saturationproblems can be solved by using infrared sensors in thesolid state and developing procedures for the acquisitionand processing of the signal in real time, making thedigital signal possible to calibrate dynamically. This tech-nology is used in the construction of some small satellitesdedicated to monitoring of fires like the IR images sensorBIRD (Bi-spectral InfraRed Detection) of DLR designed toidentify areas where fire is active (Briess et al., 2002).

Fires caused by natural and man-made events have ahuge impact on the environment, in the framework ofremote sensing activities considerable efforts are made todevelop algorithms that allow for early detection of plumesin the images, whether they are of natural or anthropogenicorigin (Chrysoulakis et al., 2007; Chung, 2002).

A large amount of smoke is emitted by fire each year.This release of smoke into the atmosphere greatly affectsthe quality of air at regional scale and can also haveimpacts on climate change (Cahoon et al. 1994). The smokeplumes travel in the atmosphere for long distances, eventhousands of kilometres, and, in some cases of particularatmospheric condition, the smoke can even reach thestratosphere. During a major accident is crucial to detectthe column of smoke from the first moments to quanti-tatively evaluate its chemical and physical composition.The analysis of smoke composition with the aid of satellitedata has an high degree of complexity since the smokedoes not have a unique spectral reflectance; the physical–chemical composition of smoke can greatly change

TSC

V

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 77

depending on combustion conditions and combustedmaterials. Usually smoke is characterized by the presenceof suspended and dissolved substances that have both adirect radiative impact, that absorbs and disperse shortwave radiation (Penner et al., 1992) and an indirectradiative impact. The smoke particles could act as con-densation nuclei of clouds and modify the optical andmicrophysical properties of clouds themselves (Kaufmanand Nakajima, 1993; Rudich et al., 2003).

Due to the difference in smoke caused by combustionconditions and materials, fires that occur from industrialaccidents emit smoke that is very different from the forestfires' smoke (Chrysoulakis and Opie, 2004); it is wellknown that during the war of 1991, the fires in Kuwaitproduced a high concentration of aerosols that had a wideparticle size distribution and a complex chemical compo-sition (Ferek et al., 1992; Hobbs and Radke, 1992; Johnsonet al., 1991; Parungo et al., 1992). The remote sensingsatellite systems are a source of useful data to analyze theplumes of smoke emitted: the location, timing (displace-ment and extension in time), the areal extension of theplume and, indirectly, an estimation of the amount ofaerosols and trace gases (Kaufman et al., 1990; Penneret al. 1992). Therefore, timely and accurate detection ofplumes has become an increasingly important issue. Thedetection of the plume caused by technological accidentshave been studied using different methods: AVHRR falsecolour images photo-interpretation (Chung and Le, 1984;Kaufman et al., 1990, Randriambelo et al., 1998, Chrysou-lakis and Cartalis 2003a, 2003b), optical thickness ofaerosol plumes (Wong and Li, 2002), methods of multi-threshold spectral indices (Baum and Trepte, 1999; Chry-

able 1ensor family used for disaster management; highlights. Modifed from IPSC –

ommission (Kucera et al., 2012).

HR¼Very High Resolution, HR¼High Resolution, MR¼Medium Resolution, SAR

soulakis and Cartalis, 2003a; Chrysoulakis et al., 2005), andmulti-threshold algorithms based on neural networks (Liet al., 2001) and methods based on the textural char-acteristics of the image (Christopher et al., 1996).

Image processing realized with MODIS and MERIS(Medium Resolution Imaging Spectrometer) techniquesare also used to detect and monitor smoke plumes and,due to their spatial and temporal resolution, have becomeimportant sources of data (Chu et al., 1998; Kaufman et al.2003; Kaufman and Tanre, 1998; King et al., 1999, Stoweet al., 1997).

Moreover air emissions that occur as "plumes" can bemonitored by their temperature difference with the sur-rounding air (Chrysoulakis et al., 2005) and often emis-sions coming from an industrial accident are composed bygases heavier than air that could lead to high concentra-tions of toxic substances at ground level. If the event hasbeen caused by a structural failure in the facility thequantity of toxic substances released and their con-centration will be extremely abundant and easily visible inremote airborne images (Dandrieux et al., 2003).

An accurate modelling of the dispersion and propaga-tion of the plume after the accident, is crucial to assess thedamage and to predict the incident scenario development:the application of an existing dispersion model (to beimplemented before the accident) will help to manage theemergency considering not only the effects in the proxi-mity of the installation, but also the impact in the neigh-boring areas and hours or days after the accident.

Kaufman and Nakajima (1993) have shown that thepresence of smoke significantly reduces the size of thedroplets in clouds, but decreases the reflectance of the

Institute for the Protection and Security of the Citizen. JRC – European

¼Synthetic Aperture Radar

Table 2EO tools used in Fukushima accident. Modified from Kazuo et al. (2012) and Iwasaki et al. (2012).

Data acquisition Platform and sensor (Kazuo et al. 2012) Optical sensor (Iwasaki et al., 2012) SAR (Iwasaki et al., 2012)

12/03/2011 – ALOS/AVNIR-2 CosmoSkymedALOS/PRISM TerraSAR-X

RADARSAT-2FORMOSTAT-2RapidEyeLANDSAT-7IKONOSSPOT-5WorldView-2ASTER (TIR)THEOS

13/03/2011 SAR Imagery (Terra SAR-X) FORMOSTAT-2 ALOS/PALSARRapidEye CosmoSkymedLANDSAT-5 TerraSAR-XGeoEye-1 RADARSAT-2SPOT-5QuickbirdEO-1

14/03/2011 Optical Imagery (EROS-B) ALOS/AVNIR-2 ALOS/PALSARRapidEye CosmoSkymed

TerraSAR-XGeoEye-1SPOT-5WorldView-1.2ASTERHJKOMPSAT-2CARTSAT-2EROS-B

15/03/2011 – ALOS/AVNIR-2 ALOS/PALSARFORMOSTAT-2 CosmoSkymedIKONOS TerraSAR-XSPOT-4 RADARSAT-2WorldView-1.2EROS-B

12/03/2011 Vertical photo (Airborne) – –

29/03/2011 Oblique photo (Helicopter) – –

05/04/2011 Mobile Mapping System (Vehicle) – –

17/03/2011 Laser Scanner (Vessel) – –

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8478

cloud because of the absorption due to the presence ofblack carbon, while Kaufman and Fraser (1997) haveshown that smoke particles increase the reflectance of thinor moderately thick clouds. This leads to considerablecomplexity in identifying plumes in remote sensing ima-ges (Shukla and Pal, 2009).

Adaktylou and Cartalis (2005) have developed a soft-ware to derive information from AVHRR images, about theplumes resulting from industrial accidents. They used as atest event the explosion that took place in a factory offireworks in the city Enschede (Netherlands) in May 2000and the huge explosion in 1991 aboard the supertankerVLCC Haven that occurred in the Gulf of Genoa.

In 2006 ESA funded the project SEVESEO (ESA, 2006SEVESEO Project; Lefebre et al., 2006) that, in the case ofindustrial accident, had the goal of providing informationon land use, topography, vegetation and atmosphericcomposition, by means of remote sensing. In this projectthe use of very high resolution data has been tested, likeQuickbird images, for detailed mapping of the accident siteas decision tool for emergency management. Land useexpeditious maps were also produced with SPOT (SatelliteProbatoire d'Observation de la Terre) images. These datawere also used to determine the surface roughness and

important parameter for modeling the spread of pollutantsreleased by the accident.

2.3. Nuclear accidents

Many authors used remotely sensed data to evaluateeffects of industrial accidents and nuclear disasters on theenvironment in the medium and long term. In this context,the best indicator of contamination is definitely thevegetation, whose spectral properties are strictly related toits physiological characteristics (state of development,health) and are therefore connected to the degree of pol-lution of the observed area. For instance the presence ofradionuclides in the soil influences the type of vegetation,affects the natural development of the plant, or inducesgenetic variation in the plant itself influencing the devel-opment of photosynthetic cells contained in the leaves(Ben-Bolie et al., 2014).

The 1986 Chernobyl incident, now an historical event,caused the contamination of a large area of Ukraine andother European regions. Immediately after the disaster,using AVHRR thermal data it was possible to detect thenuclear plant on fire by the city of Kiev even if the spatialresolution of the images (1.1 km) were not exactly ade-quate to the plant extension (Givri, 1995). In addition to

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 79

the long-lived radionuclides (137Cs, Sr 90, etc.) emittedinto the atmosphere during the incident, a huge amount ofheavy metals (Mn, Ni, Co, Cu, Pb, Zn, Th, V, Mu, YB) werereleased into the environment, and their distribution insoil can be deduced by analyzing the vegetation, in parti-cular the state of vegetation health. Already in the 80sChang and Collins (1983) and Goetz et al. (1983) hadhighlighted the shift in the blue (blue shift) of the red edgein the reflectance spectra of conifers due to contaminationwith heavy metals content in deposits of sulphides. Samemethodology was adopted by Lyalko et al. (1996) toinvestigate the presence of 137Cs in the soil after theChernobyl accident, analyzing agricultural and forestvegetation of a large area in the industrial region of wes-tern Doubas (Ukraine). In this study field data, airbornespectrometric data (spectrometer QUARTZ- 150 m alti-tude) and multispectral images taken with a KATE 200camera (4 bands, vis-nearIR, 30 m spatial resolution) andMK-4 camera (3 bands, spatial resolution 20 m, satelliteResource F2) were used.

Considering recent natech events, the largest and mostwell-known is indubitably the nuclear accident following

Fig. 4. Time-table small satellite miss

Table 3Remote sensing bands and related instruments used for oil spill detection (Ada

Band Wavelength

Radar 1–30 cmPassive microwave 2–8 mmThermal infrared (TIR) 8–14 mmMid-band infrared (MIR) 3–5 mmNear infrared 1�3 mmVisual 350–750 nmUltraviolet 250–350 nm

the earthquake and the tsunami that occurred in Japan inMarch 2011, at the Fukushima plant. The accident occurredwhen the nuclear facilities were hit by a tsunami triggeredby the magnitude 9.0 Tōhoku earthquake. On the 12nd ofMarch, radioactive material was released from the reac-tors, causing the largest nuclear incident since theChernobyl disaster in April 1986.

To analyze the territorial conditions immediately afterthe disaster and make a quick field measures to restore"normal" conditions, large number of images, were takenwith different carriers including satellites, aircraft, heli-copters, vehicles – land and naval (Yoshikawa et al., 2012,Iwasaki et al., 2012).

TerraSAR-X Satellite imagery (March 13 to April 4) wereused for the first analysis of the area affected by the tsu-nami and then they were integrated with oblique imagesof the damaged area taken by a helicopter. High resolutionoptical images were effectively used to detect damagedareas; the nuclear power plant was monitored every dayusing a constellation system of optical sensors, acquiringimages even when the cloud cover percentage was high.Considering ta some days were necessary to acquire aerial

ions (Sandau and Briess, 2010).

pted from Jha 2008).

Type of Instruments

SLAR/SARRadiometersVideo cameras and line scannersVideo cameras and line scannersFilm and video camerasFilm, video cameras and line scannersFilm, video cameras and line scanners

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8480

photo of the damaged areas, satellite data, although at alower resolution were extremely useful in the first daysduring and after the accident, before the aerial photo setwas gathered (Iwasaki et al., 2012).

Where possible, moreover, Mobile Mapping systems(land and naval means) were used to evaluate the extent ofdamage in detail (Yoshikawa et al., 2012) (Table 2).

2.4. Oil spills

Natech events include oil spills; in fact, in addition to oilspills occurring during navigation due to carelessness of oiltankers or due to “human errors”-flaws, washing tanks, tan-ker collisions, shipwrecks, platform accidents (Schmidt-Etkin,2011) oil spills may occur from refineries located near thecoast, and can be caused by natural accidents such as earth-quakes (Steinberg and Cruz, 2003) and floods (Leifer et al.,2012). To search for oil spills CleanSeaNet, a Europeandetection service, was created. The CleanSeaNet service isbased on radar satellite images, that cover all European seaareas, when an oil spill is detected, an alert message isdelivered to the relevant country (EMSA-European MaritimeSafety Agency. CleanSeaNet). Another sea and ocean mon-itored by Earth Observation (EO) project, is the SEAGOSS. Thisproject proposes advanced pattern recognition techniques toprocess remote sensing data to model sea state and oil slickdetection (de Martino et al., 2014).

The extent of damage due to spillage depends on thetype of oil, but in any case oil spill can be detected withoptical data acquired in near real time by sensors that areaboard of meteorological satellites (Grimaldi et al., 2011),even if SAR (Synthetic Aperture Radar) is the most effec-tive sensor for this purpose. In fact the oil density is lowerthan that of seawater, for this reason oil tends to remain onthe surface forming a thin film whose thickness dependson the water temperature, the composition and the natureof the oil. After that the spilled oil undergoes a series ofphysical and chemical processes (Jha et al., 2008).Such asin evaporation, oil leaves the surface, while others, such asthe generation of the water–oil emulsion, the oil persistson the surface (Brekke and Solberg, 2005). Which of thetwo types of processes prevail, depends on the specificgravity of the oil spilled. Lighter oils tend to evaporate andto dissolve rapidly (in a few hours) and most of the time donot require cleaning. Crude oil, however, breaks down anddissipates much more slowly (in a few days) requiring acleaning operation.

For this reason with remote sensing data on oil spillscan be identified observing the sea surface roughness andthis observation is commonly done with active sensor inthe wavelength range of microwaves (Jha et al., 2008)(Table 3).

Due to its high spatial resolution, the use of SAR sensorsfor the detection of small patches is essential (Ferraroet al., 2010), moreover satellite radar images provide dayand night coverage independent of fog and cloud cover(Brekke and Solberg, 2005). Such images offer the con-siderable advantage of having revisiting time of one day,allowing to continuously monitor the area where thespillage has been detected. The need for multi-temporalimages is crucial especially to monitor the movement of

the slick of lighter oil (characterized by time of persistenceon the surface of the order of a day) also detecting thedifferences between biogenic oil and mineral oil slick(Skrunes et al. 2014).

Following the DeepWater Horizon oil disaster in 2010,EO techniques were used to identify the oil spill extent,both optical, SAR and thermal. The results suggest that incase of spills, SAR data may be used to identify oil emul-sions to help make management decisions (Garcia-Pinedaet al., 2013).

However, the identification of the oil spill can often bedifficult because many natural phenomena show an elec-tromagnetic response very similar to that of the oil spill.For this reason, new optical multispectral images, withhigh spatial resolution, are merged with SAR data in orderto monitor oil spill phenomena (Jha et al., 2008).

3. Future development

Considering the previous paragraphs, between the EarthObservation services, small satellite missions can be themost useful, in case of industrial accidents when suddeninformation are needed (Kucera et al., 2012). Currentlyavailable satellite data cannot simultaneously meet theneeds of high spatial and temporal resolution that isnecessary to respond quickly in case of assisting or evalu-ating damages related to industrial accidents. The smallsatellite missions, on the other hand, could meet theserequirements, especially considering that they can bedesigned by focusing on a single topic-e.g. monitoring riskareas industrial-(Sandau 2010) using technologies alreadyavailable (off-the-shelf technologies). It is also possible tocreate a small satellite system (bus and payload) or makes abet on the engineering of the system and then on theminiaturization of the sensor component (development ofmicro-technologies for sensors). Both approaches can helpin providing relevant information in case of industrialaccidents. In particular, "small missions" (with off-the-shelftechnologies) do not require excessive budget and couldcombine the interests of both the scientific world andindustry (Sweeting 1996, 2002; Neeck and Hammer 2008).

In the case of pre and post accident monitoring thepossibility of using data acquired from specially dedicatedsensors have the advantage of a significant involvement ofsmall and medium industries, as well as a greater varietyof missions and a huge availability of application dataresulting in greater diversification of potential users.

In accordance with technological evolution, traditionalEO systems will be able to manage huge amount of data asthe cost of collection and storage decrease. Images willhave increasing spectral and radiometric resolution, asLandsat 8 images that have now a resolution of 16-bit pixelvalues (Landsat 8. http://landsat.usgs.gov/landsat8.php).Moreover the cost of hardware and imagery will decrease.Images will be processed to give information rather thanraw data. Information will be refined to the point of beingdirectly useful to risk management.

But in the near future the main effort will be madetoward producing single instrument small sats (constella-tions) to be used in specific application fields.

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 81

3.1. Small satellites

Generally, small satellites are equipped with spectro-radiometers in Vis–nearIR. The data are processed in theground station in a short time and are available after a day;in the future, it is expected the processing capacity onboard will increase and data will be sent to end usersalready processed with the corrections system (Sandau,2006).

In the next years the development of missions withsmall satellites will also be favored by the appearance onthe market of new dedicated launch systems (use ofmodified military carrier), by the need to "test" theequipment prior to organize it in a larger mission, by thedevelopment of an interconnected system of smallreceiving stations (ground station) at affordable cost and,last but not the least, by the demand for real-time data forevents with rapid evolution, such as industrial accidents ornatural disasters (Sandau, 2010, Ball, 2013).

Small satellites are easily arranged in constellationsthrough which it is possible, for example, to perform 3Dinterferometry that can be extremely useful in monitoringchanges in land use, including studying of the topographicurban-industrial deformations and estimating the dama-ges derived from the industrial accidents (Sandau andBriess, 2010) (Fig. 4).

3.2. UAV

In the future the demand will increase for optical datawith increasing resolution to be integrated with the dataacquired by the sensors operating in the microwave rangeand with the data taken from sensors mounted on aircrafts(Lewis, 2011, Sandau and Briess, 2008) or drones:

"… UAVs (Unmanned Aerial Vehicle stands for aircraftwithout human presence on board, piloted remotely froma ground station) and in particular the micro-UAVs(weighing less than 2 kg) represent the last frontier forEarth Observation at local high-resolution and low-alti-tude. Various sensors can be installed on micro-UAVs thatmake them employable in activities for land monitoring inurban and natural areas….Different types of aircrafts (air-planes, helicopters, blimps) and innovative aircraft likehelicopters multi-rotors (quadcopter and octocopter)defined as “automatic DRONS“ belong to UAV category….The experimentation has highlighted the possibility of notjeopardizing human. Recently, micro-UAVs have had aremarkable development following the increased relia-bility and reduced costs in the production of sensors basedon nano-technologies "(IUAV Department for Research,2011, various authors).

With UAVs, it is also possible to observe the Earth'ssurface with nadiral and prospective views, excellent forassessing damage due to industrial accidents. The UAVsplay a key role also in rapid mapping. For example, in thesimulation of a collapse of a school, during a trial in aproject related to smart cities, the images taken by a drone"… have produced layers of information shared on the webplatform geoSDI in very short time (in the order of tenminutes) that are definitely definable as-layers valuable-in an emergency with a very high resolution and a correct

geographic positioning even if with an error sometimes ofseveral meters" , Department IUAV for Research, 2011,Various Authors.

The data acquired with UAVs will be increasingly indemand especially for monitoring disasters both of nat-ural, anthropogenic or natech origin. Drones, in fact, cancarry on board simultaneously both cameras and instru-ments dedicated to acquire specific information about theevent to analyse. In the case of fires resulting fromindustrial accidents, for example, in which the type ofpollutants released is previously known, the sensors onthe UAV can be designed ad hoc to sample the atmosphericparticulate matter.

The possibility to use a drone would quantify themagnitude of the catastrophic event and would help topredict the area affected by the damage; it could alsoprovide information on the extent of the area on whichsmoke and ash can fall back, with obvious advantages inorganizing the procedures of intervention on the territory.

UAVs have the prerogative to be used in high-risk areassuch as those affected by major accidents of hazardouschemicals. In the case of accidents with leakage of toxicand noxious gases mini-UAV may be equipped with anautomatic platform for air monitoring to detect the con-centration of toxic substances and to collect samples to beanalysed in the laboratory (Wang et al. 2013). With opticalor infrared cameras it will be possible to monitor in realtime damages to the buildings. This can be of great supportin the classification of risk areas and evacuation routes.Using these automatic vectors, appropriately designed forthe temperatures involved and for the substances to beanalysed, allow monitoring of the areas at highest riskwhere it is impossible to access.

In the case of accidents where a chain reaction (explo-sions, fires and collapses) is expected, the use of these meansis particularly effective. In fact, in this kind of accidents anaccurate collection of information on the dynamic evolutionof the event is of fundamental importance for the formula-tion of a Search and Rescue plan and for the prevention ofdomino effect itself. The mini-UAV must necessarily beequipped with a device for detection of the temperature(thermal infrared) to operate within the accident scene (e.g.site in which the reactor is placed, unexploded tanks, inter-mediate tanks and pipelines). If the incident is not reachinghigh temperatures, a UAV could also be provided with "lifedetectors" to locate survivors, flying over areas inaccessibleto ground vehicles, supporting significantly the activities ofthe technical staff in charge of the rescue.

There are still many problems to be solved to use theUAV in case of industrial accidents, including that of theresistance to high temperatures and battery duration.

During accidents in plant containing hazardous che-micals, can be produced highly flammable gases, whichgenerate explosions. These explosions cause suddenmovements of air masses and very high temperatures. Thesmall-unmanned aerial vehicles are often not able to copewith these high temperatures; they cannot even maintainsufficient stability to minimize deformations in images.The balance, stability and control of the aircraft duringflight could be further improved (IUAV Department forResearch, 2011, various authors).

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8482

UAVs balance the lack of satellite images acquired withoptical sensors since they are able to fly under the cloudcover and are, in a certain manner, cheaper than the remotesensing monitoring system. Furthermore, the costs ofmanagement and maintenance are significantly lower thanthose of a aircraft monitoring system that need a crew.

4. Conclusions

It is well known that Earth Observation data areextremely useful to analyse the effect of industrial andnatech accidents. In the last few years, it has beendemonstrated that integrating and comparing imagestaken from different remote sensing platforms (satellites,aeroplanes and helicopters, drones) result in useful infor-mation to rescuers and the disaster management. More-over the elaboration of EO data constitutes a knowledgebase that helps to create accident scenarios, to single outrisk areas and to plan action for prevention, emergencymanagement, reconstruction and in post accident analysis.This overall view clearly shows how the EO data are asource of crucial information, both in a natech or Industrialaccident, all along the “Disaster cycle” from preparednessphase, to the emergency management, from the recon-struction back to the preparedness again in a continuousimprovement cycle.

Different spatial and spectral resolution data far frombeing only a problem have prove themselves of being a“richness” in environmental analysis, especially whenmerging procedures are well harmonized. One of thecrucial challenges to be faced in the next years will be thesmall satellites missions planning, so as to maximise theiruse in the observation of potential risk areas to improverisk assessment and management, and better understandcausality and connections between natural and man-madehazards.

Drones appear especially useful during the operationalactivities in the emergency phase, particularly if they willbe equipped with dedicated instrumentation such assensors for the detection of toxic substances, visible and IRimaging systems according to the different types of acci-dent. Drones added value consists in the possibility tocollect data simultaneously with different tools and mergethem according to event scenarios.

keeping in mind the temporal and spatial evolution ofnatech and man-made hazards, one of the main goalsseems to be prepared for a quick and efficient dataacquisition, processing and transfer. This will be the futurechallenge for spatial agencies, space industries andresearch institutions.

References

AA. VV., 2006. The Contribution of the Surveying Profession to DisasterRisk Management. FIG Publication n. 38. A publication of FIGWorkingGroup 8.4., International Federation of Surveyors (FIG), Copenhagen,Denmark.

Adaktylou, N., Cartalis, C., 2005. Detecting and monitoring plumes causedby major industrial accidents with JPLUME, a new software tool forlow-resolution image analysis. Environ. Model. Softw. 20, 1486–1494.

Alexander, D., 2002. From civil defence to civil protection and back again.Disaster Prev. Manag. 11 (3), 209–213.

Alexander, D., 2006. Crisis intervention and risk reduction RISK 21-copingwith Risks due to Natural Hazards in the 21st Century. In: Ammann,W.J., Dannenmann. S., Vulliet, L. (Eds.).

Antonioni, G., Spadoni, G., Cozzani, V., 2007. A methodology for thequantitative risk assessment of major accidents triggered by seismicevents. J. Hazard. Mater. 147 (1–2), 48–59.

Antonioni, G., Bonvicini, S., Spadoni, G., Cozzani, V., 2009. Development ofa framework for the risk assessment of Na-Tech accidental events.Reliab. Eng. Syst. Saf. 94 (9), 1442–1450.

Aschbacher, J., Milagro-Pérez, M.P., 2012. The European Earth monitoring(GMES) programme: status and perspectives. Remote Sens. Environ.120, 3–8.

Ball, M., 2013. The Small Satellite Revolution: More Earth Observations ata Lower Cost-Sensors and Systems. ⟨https://www.sensorsandsystems.com/article/features/31581-the-small-satellite-revolution-more-earth-observations-at-a-lower-cost.html#sthash.4bHb04Og.dpuf⟩.

Baum, B.A., Trepte, Q., 1999. A grouped threshold approach for sceneidentification in AVHRR imagery. J. Atmos. Ocean. Technol. 16,793–800.

Ben-Bolie, G.H., Abiama, P. E., Ateba, P.O., 2014. Impact of plant Growth inwaste-contaminated areas. In: Kumar Gupta, Walther (Eds.). Radio-nuclide Contamination and Remediation Through Plants. Springster,pp. 139–160. http://dx.doi.org/10.1007/978-3-319-07665-2_7.

Brekke, C., Solberg, A.H.S., 2005. Oil spill detection by satellite remotesensing. Remote Sens. Environ. 95, 1–13.

Briess, K., Lorenz, E., Oertel, D., Skrbek, W., Zhukov, B., 2002. Specialreport on Fire Recognition Potential of Bi-spectral InfraRed Detection(BIRD). DLR doc. n. TN-Bird-2000-WP/120.

Cahoon Jr., D.R., Stocks, B.J., Levine, J.S., III, Cofer, W.R., Pierson, J.M., 1994.Satellite analysis of the severe 1987 forest fires in northern China andsoutheastern Siberia. J. Geophys. Res. 99, 18627–18638.

Chang, S., N., Collins, W., 1983. Confirmation of airborne biophysicalmineral exploration technique using laboratory methods. Econ. Geol.78, 723–736.

Christopher, S.A., Kliche, D.V., Joyce, C., Welch, R.M., 1996. First estimatesof the radiative forcing of aerosols generated from biomass burningusing satellite data. J. Geophys. Res. 101 (D6), 21265–21273.

Chrysoulakis, N., Cartalis, C., 2003a. A new algorithm for the detection ofplumes caused by industrial accidents, on the basis of NOAA/AVHRRimagery. Int. J. Remote Sens. 24, 3353–3367.

Chrysoulakis, N., Cartalis, C., 2003b. TAD—A new satellite images analysissoftware tool for the detection of major fires caused by technologicalaccidents. Int. J. Remote Sens. 24, 1259–1271.

Chrysoulakis, N., Opie, C., 2004. Using NOAA and FY imagery to track toxicplumes caused by the 2003 bombing of Baghdad. Int. J. Remote Sens.25, 5247–5254.

Chrysoulakis, N., Adaktylou, N., Cartalis, C., 2005. Detecting and mon-itoring plumes caused by major industrial accidents with JPLUME, anew software tool for low resolution image analysis. Environ. Model.Softw. 10, 1486–1494.

Chrysoulakis, N., Herlin, I., Prastacos, P., Yahia, H., Grazzini, J., Cartalis, C.,2007. An improved algorithm for the detection of plumes caused bynatural or technological hazards using AVHRR imagery. Remote Sens.Environ. 108, 393–406.

Chu, D.A., Kaufman, Y.J., Remer, L.A., Holben, B.N., 1998. Remote sensingof smoke from MODIS airborne simulator during the SCAR-Bexperiment. J. Geophys. Res. 103, 31979–31988.

Chung, Y.S., 2002. Satellite observations of smoke plumes from forest firesin Canada. Int. J. Remote Sens. 23, 2341–2343.

Chung, Y.S., Le, H.V., 1984. Detection of forest-fire smoke plumes bysatellite imagery. Atmos. Environ. 18 (10), 2143–2151.

Clerc, A., Le Claire, G., 1994. The environmental impacts of natural andtechnological (Na-tech) disasters. Background Discussion Paper forThe World Conference on Natural Disaster Reduction, Yokohama,Japan, 23–27 May.

Cruz, A.M., Steinberg, L.J, Vetere Arellano, L., Nordvik, J.P., Pisano, F., 2004.State of the Art in natech Risk Management (NATECH: Natural HazardTriggering a Technological Disaster). EUR 21292 EN, © EuropeanCommunities.

Dandrieux, A., Dusserre, G., Thomas, O., 2003. The DVS model: a newconcept for heavy gas dispersion by water curtain. Environ. Model.Soft. 18, 253–259.

de Martino, M., Dellepiane, S., Gemme, L., Moser, G., Serpico, S.B.,Toma, M., Degano, C., Loiaconi, A., Mainenti, I., Cusati, L.A., Pedroncini,A., 2014. In: Geoscience and Remote Sensing Symposium (IGARSS),IEEE International. pp. 4454–4457. http://dx.doi.org/10.1109/IGARSS.2014.6947480.

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–84 83

Dyke, G., Gill, S., Davies, R., Betorz, F., Andalsvik, Y., Cackler, J., Dos Santos,W., Dunlop, K., Ferreira, I., Kebe, F., Lamboglia, E., Matsubara, Y.,Nikolaidis, V., Ostoja-Starzewski, S., Sakita, M., Verstappen, N., 2010.Dream project: Applications of Earth Observations to disaster riskmanagement. Acta Astronaut. http://dx.doi.org/10.1016/j.actaastro.2010.06.018.

El Hajji, C., Piatyszek, E., Laforest, V., 2013. Development of generic sce-nario of industrial accidents triggered by floods: a first step towarddecreasing the vulnerability of industrial facilities. Risk analysis VIII,in: Brebbia, C.A. (Ed), Risk analysis VIII, Witt press.

Ehrlich, D., Tenerelli, P., 2013. Optical satellite imagery for quantifyingspatio-temporal dimension of physical exposure in disaster riskassessments. Nat. Hazard. 68, 1271–1289, http://dx.doi.org/10.1007/s11069-012-0372-5.

ESA-European Space Agency. SEVESEO Project2006. ⟨http://www.seveseo.eu/product/Pages/default.aspx⟩.

European Commission-Joint Research Centre (JRC), 2004. Analysis ofnatech (Natural Hazard Triggering Technological Disasters) DisasterManagement. In: Vetere Arellano, A.L., Cruz, A.M., Nordvik, J., Pisano,F., NEDIES Workshop Proceedings Ispra, Italy, 20–21 October 2003.

Ferek, R.J., Hobbs, P.V., Herring, J.A., Larsen, K.A., Weiss, R.A., 1992. Che-mical composition of emissions from Kuwait oil fires. J. Geophys. Res.97 (D13), 14483–14489.

Ferraro, G., Baschek, B., de Montpellier, G., Njoten, O., Perkovic, M., Vespe,M., 2010. On the SAR derived alert in the detection of oil spillsaccording to the analysis of the EGEMP. Mar. Pollut. Bull. 60, 91–102.

French Ministry of Ecology, Sustainable Development and Energy. Townplanning, pollution and risk prevention Department. Appendix2_foreign accidents. ARIA (analysis, research and information onaccidents) database.

Galderisi, A., Ceudech, A., Pistucc, M., 2008. A method for Na-tech riskassessment as supporting tool for land use planning mitigationstrategies. Nat. Hazard. 46 (2), 221–241. 10.1007/s11069-008-9224-8.

Garcia-Pineda, O., MacDonald, I., Hu, C., Svejkovsky, J., Hess, M.,Dukhovskoy, D., Morey, S.L., 2013. Detection of floating oil anomaliesfrom the Deepwater Horizon oil spill with synthetic aperture radar.Oceanography 26 (2), 124–137.

Girgin, S., Krausmann, E., 2013. RAPID-N: Rapid natech risk assessmentand mapping framework. J. Loss Prev. Process Industries 26, 949–960,http://dx.doi.org/10.1016/j.jlp.2013.10.004.

Givri, J.R., 1995. Satellite remote sensing data on industrial hazards. Adv.Space Res Vol. 15 (Issue 11), 87–90, http://dx.doi.org/10.1016/0273-1177(95)00078-S.

Goetz, A.F., Book, B.N., Rowan, L.C., 1983. Remote sensing for exploration.Econ. Geol. 78, 573–590.

Grimaldi, C.S.L., Casciello, D., Coviello, I., Lacava, T., Pergola, N., Tramutoli, V.,2011. An improved RST approach for timely alert and Near Real Timemonitoring of oil spill disasters by using AVHRR data. Nat. Hazards EarthSyst. Sci. 11, 1281–1291.

Hobbs, P.V., Radke, L.F., 1992. Airborne studies of the smoke from theKuwait oil fires. Science 256, 987–991.

Iwasaki, A., Miyatami, S., Nakasuka, S., 2012. Satellite contributions to disastermonitoring- Japanese Earthquake and Tsunami Case in 2011. In: 26thAnnual AIAA/ASU Conference on Small Satellites, SSC12-XII-1.

Jha, M.N., Levy, J., Gao, Y., 2008. Advances in remote sensing for oil spilldisaster management: State-of-the-art sensors technology for oil spillsurveillance. Sensors 8, 236–255.

Johnson, D.W., Kilsby, C.G., Mc Kenna, D.S., Saunders, R.W., Jenkins, G.L.,Smith, F.B., 1991. Airborne observations of the physical and chemicalcharacteristics of the Kuwait oil smoke plume. Nature 353, 617–621.

Joyce, K.E., Wright, K.C., Samsonov, S.V., Ambrosia, V.G., 2009. Remotesensing and the disaster management cycle. In: Gary Jedlovec (Ed.),InTech, ISBN: 978-953-307-005-6 http://dx.doi.org/10.5772/8341.

Kaufman, Y.J., Tucker, C.J., Fung, I., 1990. Remote sensing of biomassburning in the tropics. J. Geophys. Res. 95, 9927–9939.

Kaufman, Y.J., Nakajima, T., 1993. Effect of Amazon smoke on cloudmicrophysics and albedo – Analysis from satellite imagery. J. Appl.Meteorol. 32, 729–744.

Kaufman, Y.J., Fraser, R.S., 1997. The effect of smoke particles on clouds,and climate forcing. Science 277, 1636–1639.

Kaufman, Y.J., Tanre, D., 1998. Algorithm for remote sensing of tropo-spheric aerosol from MODIS: Algorithm theoretical basis document.Greenbelt, Md. NASA Goddard Space Flight Cent.Greenbelt, Md.

Kaufman, Y.J., Ichoku, C., Giglio, J., Korontzi, S., Chu, D.A., Hao, W.M., 2003.Fire and smoke observed from the Earth Observing System MODISinstrument—Products, validation, and operational use. Int. J. RemoteSens. 24, 1765–1781.

King, M.D., Kaufman, Y.J., Tanre, D., Nakajima, T., 1999. Remote sensing oftropospheric aerosols from space: Past, present, and future. Bulletinof the American Meteorological Society, 11, pp. 2229�2259.

Krausmann, E., Cozzani, V., Salzano, E., Renni, E., 2011. Industrial acci-dents triggered by natural hazards: an emerging risk issue. Nat.Hazard. Earth Syst. Sci. 11, 921–929, http://dx.doi.org/10.5194/nhess-11-921-2011.

Krausmann, E., Cruz, A.M., 2013. Impact of the 11 March 2011, Great EastJapan earthquake and tsunami on the chemical industry. Nat. Hazard.67, 811–828, http://dx.doi.org/10.1007/s11069-013-0607-0.

Kucera, J., Lemoine, G., Kemper, T., 2012. Post-Disaster Needs Assessment:the role of remote sensing and geospatial information. Vienna. IPSC-Institute for the Protection and Security of the Citizen. JRC- EuropeanCommission (17 February 2012, PDNA Training).

Lefebre, F., Meuleman, K., Sterckx, S., Gontier, E., Paganini, M., 2006.SEVESEO Industrial Disaster Management from Space. ESA fundedEuropean project.

Leifer, I., Lehr, W.J., Simecek-Beatty, D., Bradley, E., Clark, R., Dennison, P.,Hu, Y., Matheson, S., Jones, C.E., Holt, B., Reif, M., Roberts, D.A., Svej-kovsky, J., Swayze, G., Wozencraft, J., 2012. State of the art satelliteand airborne marine oil spill remote sensing: application to the BPDeepwater Horizon oil spill. Remote Sens. Environ. 124, 185–209,http://dx.doi.org/10.1016/j.rse.2012.03.024.

Lewis, P.E., 2011. The Evolution of Airborne Chemical and RadiologicalRemote Sensing For Emergency and Natural Disaster Response.National Geospatial-Intelligence Agency (NGA)-USA. SPIE 2011Remote Sensing PlenaryTalk; ⟨http://spie.org/Documents/AboutSPIE/PDF/ERS11-plenary-Lewis.pdf⟩.

Li, Z., Khananian, A., Fraser, R.H., Cihlar, J., 2001. Automatic detection offire smoke using artificial neural networks and threshold approachesapplied to AVHRR imagery. IEEE Trans. Geosci. Remote Sens. 39,1859–1870.

Lindell, M.K., Perry, R.W., 1996. Identifying and managing conjointthreats: earthquake-induced hazardous materials releases.US J. Hazard. Mater. 50, 31–46.

Lyalko, V.I., Djary, V.Y., Sakhatsky, A.I., Hodorovsky, A.J., Woolfson, L.D.,Shportjuk, Z.B., Sibirtseva, O.N., 1996. Estimation of heavy metal andradionuclide contamination of soils and vegetation within the Cher-nobyl danger zone using remote sensing data. Int. Arch. Photogramm.Remote Sens., XXXI; .

Marzo, E., Busini, V., Rota, R., 2012. Definition of a shortcut methodologyfor assessing flood-related Na-Tech risk. Nat. Hazard. Earth Syst. Sci.12, 3241–3253, http://dx.doi.org/10.5194/nhess-12-3241-2012.

Marzo, E., Busini, V., Rota, R., 2015. Definition of a short-cut methodologyfor assessing the vulnerability of a territory in natural-technologicalrisk estimation. Reliab. Eng. Syst. Saf. 134, 92–97.

Menoni, S., Parker, D., Ballio, F., Tapsell, S., 2012. Assessing multifacetedvulnerability and resilience in order to design risk-mitigation stra-tegies. Nat. Hazard. 64, 2057–2082, http://dx.doi.org/10.1007/s11069-012-0134-4.

Neeck, S.P., Hammer, T.F., 2008. Nasa Earth Observation Programs andsmall satellites. Small Satellites for Earth Observation.

Parungo, F., Kopcewicz, B., Nagamoto, C., Schnell, R., Sheridan, P., Zhu, C.,1992. Aerosol particles in the Kuwait oil fire plumes—Their mor-phology, size distribution, chemical composition, transport, andpotential effect on climate. J. Geophys. Res. 97, 15867–15882.

Penner, J.E., Dickinson, R.E., O'Neill, C.A., 1992. Effects of aerosol frombiomass burning on the global radiation budget. Science 256,1432–1434.

Randriambelo, T., Baldy, S., Bessafi, M., Petit, M., Despinoy, M., 1998. Animproved detection and characterization of active fires and smokeplumes in south-eastern Africa and Madagascar. Int. J. Remote Sens.19, 2623–2638.

Rout, D.K., Parida, P.K., Behera, G., 2005. Man-made disaster a case studyof Nalco ash-pond in the Angul district, Orissa, using remote sensingand gis technique. Photonirvachak. Joumal of the Indian Society ofRemote Sensing 33 (2).

Rudich, Y., Sagi, A., Rosenfeld, D., 2003. Influence of the Kuwait oil firesplume (1991) on the microphysical development of clouds. J. Geo-phys. Res. 108, 4478, http://dx.doi.org/10.1029/2003JD00347.

Salzano, E., Basco, A., Busini, V., Cozzani, V., Marzo, E., Rota, R., Spadoni, G.,2013. Public awareness promoting new or emerging risks: Industrialaccidents triggered by natural hazards (NaTech). J. Risk Res. 16 (3–4),469–485, http://dx.doi.org/10.1080/13669877.2012.729529.

International Study on Cost-Effective Earth Observation Missions. In:Sandau, R. (Ed.), 160. , Balkema A.A. Publishers, Taylor & FrancisGroup, Leiden, The Netherlands. ISBN10:0-415-39136-9, isbn13:9-78-0-415-39136-8.

S. Di Franco, R. Salvatori / Remote Sensing Applications: Society and Environment 1 (2015) 72–8484

Sandau, R., Briess, K., 2008. Potential for advancements in remote sensingusing small satellites. The International Archives of the Photo-grammetry, Remote Sensing and Spatial Information Sciences,XXXVII.

Sandau, R., Briess, K., 2010. The role of small satellite mission in globalchange studies- Advances. In: Chuvieco, E., Li, J., Yang, X. (Eds.),Springer Science & Business Media, pp. 298.

Sandau, R., 2010. Status and trends of small satellite missions for EarthObservation. Acta Astronaut. 66, 1–12.

Schmidt Etkin, D., 2011. Spill Occurrences: A World Overview. Chapter 2.Oil spill science and technology: prevention, response, and cleanup.In: Fingas, M. (Ed.), Elsevier, pp. 7–48, http://dx.doi.org/10.1016/B978-1-85617-943-0.10002-4.

Sengupta, A., 2007. Industrial hazard, vulnerability and Risk Assessmentfor land use Planning: A case study old Haldia, west Bengal, India.Thesis submitted to the International Institute for Geo-informationScience and Earth Observation (ITC). ⟨http://www.itc.nl/library/papers_2007/msc/iirs/sengupta.pdf⟩.

Showalter, P.S., Ramspott, M., 1999. The use of remote sensing indetecting and analyzing natural hazards and disaster, 1972–1998. Apartially annotated Bibliography. Published by the J.M. Lovell Centerfor Environmental Geography and Hazards Research, SouthwestTexas State University. Paper N.1.

Shukla, B.P., Pal, P.K., 2009. Automatic smoke detection using satelliteimagery: preparatory to smoke detection from Insat-3D. Int.J. Remote Sens. 30 (1), 9–22.

Skrunes, S., Brekke, C., Eltoft, T., 2014. Characterization of marine surfaceslicks by radarsat-2 multipolarization features. IEEE Trans. Geosci.Remote Sens. 52 (9), 5302–5319, http://dx.doi.org/10.1109/TGRS.2013.2287916.

Steinberg, L.J., Cruz, A.M., 2003. When Natural and Technological Dis-asters Collide: Emergency Management Lessons From the TurkeyEarthquake of August 17, 1999. Nat. Hazard. Rev. 5 (3), 121–130.

Stowe, L.L., Kaufman, Y.J., Herman, M., Matoo, S., 1997. Remote sensing ofaerosol properties over oceans using the MODIS/EOS spectralradiances. J. Geophys. Res. 102, 16971–16988.

Sweeting, M.N., 1996. Why satellites are scaling down, Space TechnologyInternational. p. 55–59, in: Sandau, R. (Ed.), International Study onCost-Effective Earth Observation Missions, Balkema, A.A. Publishers, amember of Taylor & Francis Group plc, Leiden, The Netherlands, 2006,p. 160. ISBN 10: 0-415-39136-9, ISBN 13: 9-78-0-415-39136-8.

Sweeting, M.N., 2002. Micro/Nanosatellite-the new World. In: Reycroft,M., Crosby, N. (Eds.), Kluwer Academic Publishers, Netherlands,pp. 1–19.

Van Westen, C.J., 2002. Remote sensing and geographic informationsystems for natural disaster management. In: Skidmore, Andrew (Ed.),Taylor and Francis.

Wang, L., Zhou, W., Zhao, S., 2013. Application of Mini-UAV in EmergencyRescue of Major Accidents of Hazardous Chemicals. In: InternationalConference on Remote Sensing. Environment and TransportationEngineering (RSETE 2013).

Wong, J., Li, Z., 2002. Retrieval of optical depth for heavy smoke aerosolplumes: Uncertainties and sensitivities to the optical properties.J. Atmos. Sci. 59, 250–261.

Yoshikawa, K., Okajima, Y., Takagishi, S., 2012. Disaster Monitoring usingremote sensing for the Great East Japan Earthquake. TS03K - SpecialSession on Catastrophic Disaster of East Japan Earthquake and Tsu-nami. FIG Working Week 2012. Knowing to manage the territory,protect the environment, evaluate the cultural heritage, Rome, Italy,6–10 May 2012-1/8.